Hello,
I've subscribed to the list for quite a while, and I sincerely hope this is the
correct place to report this particular problem. If not, I apologize and ask
that someone point me in the correct direction.
Since I would rather include too much information than too little, I'll include
a brief description of the hardware, the distro, and the problem, followed by a
printout of the oops.
The machine:
SMP dual PII/400
Supermicro P6DBU
256MB PC-100 in 4 sticks of 64MB ea.
Adaptec 2940UW
TNT2 Ultra 32MB AGP
(2) Intel PRO/100 PCI NICs
PCI es1370 sound
LinTV card (bttv)
300w Antec PSU
APC 500 PS backup
Mandrake 7.1, expert install, chose everything, and went with ReiserFS (love
those speedy boots). This box ran very well for weeks. The box itself is
hand-built, and has run well for a year or so under various Linux distros.
When 2.2.17 was released, I immediately went and pulled the 2.2.17 kernel, and
the 3.5.24 reiserfs patch. Patched, config'd, built. Started having terrible
lockups. Closing Netscape would do it almost every time. I removed all Netscape
from the machine. Mozilla (latest nightly) would crash it on opening up a child
window. I started having massive corruption of data streams. For instance, when
opening up *.tar.gz files, or checking RAR archives, etc. md5 would come up with
inconsistent results (!). Files corrupted (but only ones that had been
"touched"/manipulated. Samba would happily take down the machine. All of these
are hard locks, requiring manual reboots.
Rebuilt the kernel several times. Pulled fresh source, fresh patch (now at
3.5.26), reapplied, copied old .config over and did 'make oldconfig', and on
others just simply redid the config from scratch. All of these in just about
every way there is to arrange 9 books on a shelf. Chose modules, chose built-in,
on and on ad nauseum. I surely did not want to post to this list without at
least doing my homework.
I have other mission-critical boxen that I have held off of the 2.2.17 upgrade
until I'm comfortable that this is just a problem specific to this machine or my
stupidity. This is the only SMP box I have, the others are UP. The following was
taken from my /var/messages log.
I appreciate any/all feedback, because this is just quite simply driving me
crazy. Most un-Linuxish behavior. :(
>Oct 2 11:22:58 omega kernel: <3>kmem_free: Bad obj addr (objp=ccc58280, name=size-32)
>Oct 2 11:22:58 omega kernel: <1>Unable to handle kernel NULL pointer dereference at virtual address 00000000
>Oct 2 11:22:58 omega kernel: <1>current->tss.cr3 = 0f533000, %%cr3 = 0f533000
>Oct 2 11:22:58 omega kernel: <1>*pde = 00000000
>Oct 2 11:22:58 omega kernel: <4>Oops: 0002
>Oct 2 11:22:58 omega kernel: <4>CPU: 1
>Oct 2 11:22:58 omega kernel: <4>EIP: 0010:[kfree+410/460]
>Oct 2 11:22:58 omega kernel: <4>EFLAGS: 00010202
>Oct 2 11:22:58 omega kernel: <4>eax: 00000039 ebx: cffff080 ecx: 0000004a edx: 00000002
>Oct 2 11:22:58 omega kernel: <4>esi: ccc58280 edi: 00000202 ebp: ffffffff esp: ca955ecc
>Oct 2 11:22:58 omega kernel: <4>ds: 0018 es: 0018 ss: 0018
>Oct 2 11:22:58 omega kernel: <4>Process smbd (pid: 624, process nr: 31, stackpage=ca955000)
>Oct 2 11:22:58 omega kernel: <4>Stack: 00000000 ffffffff ccd58b00 ca682860 a54eebaa 00000011 cc29b404 ca955f68
>Oct 2 11:22:58 omega kernel: <4> 00000040 c0136071 ccc58280 ca682860 ca71001a 00000000 ccd58b00 ca58c8a0
>Oct 2 11:22:58 omega kernel: <4> fffffffe ca710000 08124d60 bffff75c 00000000 ffffffff ca8e4e40 ca955f78
>Oct 2 11:22:58 omega kernel: <4>Call Trace: [prune_dcache+229/312] [shrink_dcache_parent+23/48] [lookup_dentry+365/504] [vfs_rmdir+254/308] [sys_rmdir+201/320] [common_interrupt+24/32] [system_call+52/64]
>Oct 2 11:22:58 omega kernel: <4> [stext+43/164]
>Oct 2 11:22:59 omega kernel: <4>Code: c7 05 00 00 00 00 00 00 00 00 eb 1d 89 f6 83 c4 f8 56 68 82
-- Brad Felmey - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Oct 07 2000 - 21:00:10 EST